PURPOSE: The purpose of this study is to compare the effects of HOE 901 and NPH on glycated hemoglobin and to compare the safety of HOE 901 with NPH in subjects with type I diabetes mellitus. A comparison between the two treatments will also be made in terms of blood glucose variability, hypoglycemia, other indicators of metabolic control, quality-of-life, and pharmacoeconomics. METHODS: This is a phase III, randomized, multicenter, open, NPH human insulin controlled, randomized (1:1), 28-week parallel-group study with two treatment groups (HOE901 and NPH human insulin). A total of 520 patients ( a total for all sites) will be evaluated in this study. The study consists of a 1 to 4-week screening phase and a 28-week treatment phase which includes an initial active dose titration phase. Subjects randomized to the NPH insulin group will continue their previous regimen of injections per day. Subjects randomized to HOE901 will receive a single injection of HOE 901 at bedtime. Both treatment groups will also receive regular insulin in addition to either HOE901 or NPH human insulin. The subjects will be stratified by whether they were being treated with a basal insulin once versus twice daily. This study will involve 5 outpatient visits and 3 inpatient visits. During the screen visit, subjects will undergo a history and physical examination, blood tests, pregnancy screen, dilated eye exam, and fundus photography. Subjects will be instructed in home blood glucose monitoring, issued a glucose meter and supplies, and will be asked to monitor their blood glucose 4x/day. At the following visit, patients who qualify will be admitted to Rankin Clinical Research Unit for approximately 36 hours, during which 24 hr blood glucose samples will be taken. Patients will also have an EKG, blood tests, and will be asked to complete 2 questionnaires about quality of life (QOL) and to answer several pharmacoeconomic questions. Pharmacoeconomics will be assessed throughout the treatment phase in terms of direct costs (health care resource utilization) and indirect costs (lost time from work). Patients are randomized at this visit. Phone contact between patient and investigator will occur at least twice each week during dose titration. Outpatient follow-up visits including blood tests will occur at Weeks 1 & 4. Patients will be readmitted for an inpatient stay, including 24 hr sampling, at Week 8. Outpatient follow-up visits including blood tests will occur at Weeks 12 & 20, with optional eye exam and fundus photos at Week 12. The final admission will occur at Week 28. In addition to 24 hr sampling, an EKG, eye exam, fundus photos, and blood tests will be performed. Patients will restart their prestudy insulin regimen at this visit. Patients will complete the QOL questionnaires again at Weeks 8 & 28.